SK hynix Inc. (A000660) Earnings Call Transcript & Summary

[Interpreted] Good morning and thank you all for joining us today. We will now begin SK hynix' 2023 First Quarter Earnings Conference Call. [Operator Instructions] A company presentation will be simultaneously translated and the Q&A session will be consecutively translated. With that, we are now ready to begin.

[Interpreted] Good morning, and good afternoon and evening to those calling in from abroad. This is Park Seong Hwan, the Head of IR at SK hynix. Welcome to the SK hynix 2023 First Quarter Earnings Release Conference Call. Before starting the conference call, allow me to introduce the executives present here with me today. First, Kim Woo-Hyun, CFO; Park Myoung-Soo, Head of DRAM Marketing; and Park Chan-dong, Head of NAND Marketing. Let me issue a disclaimer that all outlooks presented by our company are subject to change depending on the macroeconomic and market circumstances. With that, we will now begin SK hynix' earnings release conference call. Mr. Kim will first present the earnings for first quarter 2023, followed by the company's plan and market outlook. There will then be a Q&A session with the executive.

[Interpreted] Good morning. Allow me to first report to you the company's performance of first quarter 2023. First quarter saw continued weakness in the semiconductor demand due to financial market instability and concerns of global economic downturn. With seasonality on this ongoing customers' inventory adjustments, DRAM and NAND shipments fell significantly compared to the previous quarter, and prices continued to decline across all applications. Accordingly, our first quarter revenue recorded KRW 5.09 trillion, down 34% from the previous quarter and down 58% year-on-year. For DRAM, bit shipments fell by around 20% quarter-on-quarter, while ASP drop has been moderated to fall by high-teen percent. NAND shipment growth fell by mid-teen percent quarter-on-quarter and ASP dropped by around 10%. The rate of ASP decline was more moderate compared to the market as a result of conservative sales in response to the rapidly deteriorating NAND market conditions. First quarter costs of [ sold ] decreased due to lower bit sales and SG&A also fell from cost reduction efforts. However, a rapid decline in revenue as well as increased inventory valuation loss from higher inventory volumes and further ASP decline led to first quarter operating loss to record KRW 3.40 trillion and operating margin of negative 67%. Depreciation and amortization in the first quarter was KRW 3.56 trillion, slightly lower than that of the previous quarter as a result of reduced capital spending. EBITDA was KRW 0.15 trillion and EBITDA margin was 3%. Total nonoperating loss net of gain was KRW 0.12 trillion. This includes net interest expense of KRW 0.25 trillion as well as foreign currency-related net gain of KRW 0.1 trillion. Net loss before tax was KRW 3.53 trillion, and net loss for the quarter was KRW 2.59 trillion, with a net profit margin of negative 51%. Consolidated cash balance at the end of Q1 was KRW 6.14 trillion, down by KRW 0.27 trillion from the end of last year. Total interest-bearing debt was KRW 28.76 trillion, which is an increase by KRW 5.76 trillion from a quarter ago. The company's debt-to-equity and net debt-to-equity ratio at the end of the quarter was 47% and 37%, respectively, up from 36% and 26% at the end of last year. I will now move on to the company's market outlook and future plans. Amidst lingering global economic uncertainties, memory demand is expected to decrease from expectations earlier in the year. However, customers' memory inventory levels generally declined throughout first quarter. And inventory across the memory industry are expected to improve from the second quarter with suppliers' production cuts taking into full effect. IT device shipments this year are expected to remain weak, and thus, memory demand growth this year will be driven by content growth. With memory price that has fallen by more than 60% since peak level, increase in memory contents on back of price elasticity is expected to drive demand. Demand recovery driven by content growth and the effects from supply reduction will bring a more balanced situation for demand and supply as we move forward into the remaining half of the year. Furthermore, next year's industry production growth is expected to be limited due to reduced production capabilities following CapEx cuts this year. Consequently, the strength of recovery in the coming upturn cycle will be significant, considering the severity of the current downturn. Looking at demand by application. PC shipment is expected to decline for 2 consecutive years. Weakness in consumer demand is aggravated by a reduced purchasing power due to inflation, as well as higher service consumption since reopening. However, memory contents are increasing with higher sales of high-spec laptops and gaming PCs. While it is also difficult to expect growth in smartphone unit sales this year, inventories are being reduced at some distribution channels, and memory inventories at Chinese mobile customers are also in declining trend. In the second half, demand is expected to recover with an economic rebound in China, together with the launch of new products. Smartphone demand remains polarized between flagship and low to mid-end models. And memory contents growth is witnessed in flagship models which have lower price sensitivity. In addition, certain Chinese customers have been showing interest in high-density products of 16 gigabyte or higher LPDDR5X and high-performance products, such as 9.6-gigabit per second turbo in order to expand sales through product differentiation. These differentiation efforts are expected to drive mobile memory demand this year. Demand in the server market is expected to soften this year due to tightened corporate IT spending and inventory adjustments by CSP customers. However, server customers' memory inventory is also in declining trend. And since the majority of the inventory is DDR4 products, this year's demand is expected to rapidly shift towards DDR5. In addition, demand for 128 gigabyte or higher-density DDR5 modules is quickly rising as competition intensifies for technology development and commercialization of large language models and generative AI such as ChatGPT. And as a result, would stimulate server memory demand this year. All in all, memory demand for this year is projected to grow by mid- to high single digit for DRAM and mid- to high-teen for NAND with potential upside demand in the second half depending on macroeconomic situation. We will flexibly respond to meet customers' demand by actively shifting product mix as required. Our sales bit growth in the second quarter is expected to grow by double-digit sequentially for both DRAM and NAND. This has more than offset the level of the decline we saw in the first quarter. Not limited to the base effect of low sales volume in the previous quarter, we are also planning to proactively respond to growing market demand of DDR5 and LPDDR5 as well as 176-layer SSD and uMCP products. Therefore, we expect meaningful revenue growth for the second quarter. Memory markets saw an unprecedentedly low level of demand growth last year. And demand growth outlook for this year is again being revised down from earlier forecast. In particular, the DRAM industry, which had been generating profits despite market fluctuations and evidently turned to a loss in the first quarter. In response to these [ tire ] market conditions, we are executing our CapEx in line with our previous plans to cut back spending by more than 50% on a consolidated basis compared to that of last year. This implies minimized level of investments in every aspect, excluding the essential investments to maintain competitiveness. However, we are maintaining investments for products such as DDR5, LPDDR5 and HBM3 which are expected to drive demand this year. Also by securing readiness for mass production of 1B nanometer DRAM and 238-layer NAND within the year, we are planning to be fully prepared for swift customer support when market conditions improve. In efforts to bring down excessive inventory and to bring forward the timing of market balance, we reduced wafer inputs of some legacy and low-margin products during the fourth quarter, and the effects from the lower wafer loadings have been gradually materializing from the first quarter. However, reflecting the current outlook on this year's market demand, which has been lowered from prior expectations and in order to flexibly operate our production facilities in line with the market demand and profitability, we are now adjusting wafer starts of products that have relatively high inventory levels. Now that all manufacturers are cutting back production and as the impact from adjustments start to materialize more meaningfully in the second quarter, we expect inventory for the industry to reach a more normalized level within this year. Despite the difficult market conditions today, we'll maintain our competitiveness by focusing on technology and product development, which will be the driving force for future growth. The proportion of our core products of 1A nanometer DRAM and 176-layer NAND out of total production is not likely to be expanded at a normal pace due to the CapEx cuts. However, yield rates of all products using 1A nanometer DRAM and 176-layer NAND have now reached very stable levels and, therefore, will drive cost reductions this year. Furthermore, we will maintain industry-leading technological competitiveness by completing ramp operations of next-generation 1B nanometer and 238-layer product by mid-year. Compared to 1A nanometer, 1B nanometer products net die efficiency is significantly improved, while process efficiency is also maximized with increased level of EUV applications. Having already secured high yield rates in initial testing stage, we will be able to quickly shift to these products once market conditions improve next year. Growth of DDR5 market in which we have a competitive advantage is expected to accelerate this year. Our DDR5 products, which were introduced early in the industry, have the most number of products that are validated by CPU companies. In addition, as the only supplier in the industry that provides 1A nanometer-based high-density models -- modules such as 128 gigabytes or more, we are well positioned to effectively support customers demand that require intensive memory workloads such as AI. Moreover, we have also completed the development of 24 gigabytes HBM3, which is the largest density currently available, and the product is now being shipped for customer samples. By stacking 12 layers of chips using TSV technology for the first time in the industry, we have maintained the product height same as the 16-gigabyte product, which adds to our product competitiveness. Leveraging such product competitiveness, we are anticipating our sales of 128-gigabyte and higher-density DDR5 DRAM modules this year to increase by more than 6x while sales of HBM products to increase by more than 50% this year compared to that of last year. We expect the growth of sales for these products to continue in the full following year. Having the competitive edge, such as 1A nanometer-based DRAM full product lineup and far outpacing market position in HBM, we will absorb growing AI demand in the future and continue to solidify our leadership position in the industry. Last but not least will be the company's ESG management activities. Last July, the company disclosed mid-term ESG targets for 2030 based on the ESG framework called PRISM and is implementing ESG management in steps by annually reviewing yearly goals and their progress. At the ESG Management Committee held in early April, the performance of the past year was reviewed and as a result have found 21 out of 27 annual goals were achieved. This includes several goals of high interest, such as 100% rate of renewable energy consumption in global sites. For areas that have insufficient results, the ESG Management Committee led by the company's key executives will continuously evaluate and make efforts for improvements. Despite the rapidly deteriorating memory market and continuing uncertainties throughout all areas of our business, we will strive to efficiently achieve our mid- to long-term targets for 2030 by progressively promoting ESG management based on PRISM framework. Thank you. With that, we are now ready to take your questions.

[Foreign Language] [Operator Instructions] [Foreign Language] The first question will be presented by Ricky Seo from HSBC.

[Interpreted] First, you mentioned that -- actually you provided a softer demand outlook compared to the earlier forecast. I want to know what exactly you're seeing across applications that has changed that drives this softer outlook. And then secondly, what inventory levels are you seeing? Within SK hynix, what are your DRAM and NAND inventory levels as well as the breakdown across customers as you see it? Specifically if you can provide a quarter-on-quarter comparison, that would be very helpful.

[Interpreted] Let me first address your first question on the slowdown in demand. Indeed, this year, we expected that there would be stronger growth compared to the prior year as falling ASPs would likely drive memory content growth as well as we're seeing some progress in the inventory digestion at customers. And also there was the expectation for China's reopening to translate into an economic recovery. But now what we're seeing is continued macroeconomic uncertainties as well as a further slowing of consumer sentiment that now drives our newer outlook for mid- to high single-digit growth in DRAM and mid- to high double-digit -- excuse me, mid- to high teens growth for NAND. So again, we're seeing slower shipment growth across applications in our newer outlook compared to early year, and we now expect a decline across all applications year-over-year for server, PC and smartphones. That said, we continue to believe that content growth driven upside as well as the move towards high densities, higher capacities driving demand remains intact with our original outlook. Judging from the movement of the past few years, we believe that this first quarter would represent the bottom for such shipments for our end applications. And as we move into the second half of the year, we expect some improvement over the first half. Across applications for the smartphone segment, we believe that the key point in the second half, the key driver would again be the movement towards higher capacity models that we are seeing and hearing from our customers. And then for the PC segment, although the situation overall is quite dim, we expect there to be some replacement demand even after the special upside that we had seen throughout the pandemic with the upside driven by a Chromebook demand and also the various value experiences that were required during that remote environment. On top of that, as was briefly mentioned in the remarks, generative AI models are driving fast growth in demand for the higher density and high bandwidth products. We believe this would be an upside factor into this year as well as the next.

[Interpreted] Thank you for your question. I will take your question on Hynix' inventory level as well as the expected peak as we see it. In the first quarter, despite production cuts due to the significant decline in revenue, we saw finished goods inventory grow for both DRAM and NAND. That said, as we mentioned in the prepared remarks, we are currently adjusting production primarily for the highly stocked products and also expect the sales in second quarter to more than offset in terms of growth, the decline we saw in the first quarter. All in all, we expect this to lead our inventory levels to peak in the first half and gradually decline thereafter. Now about customer inventory levels as we see it. Overall, we are seeing a decline across customers but still some elevated levels at particular customers in particular applications. We believe that the elevated inventory levels that remain at some customers, as mentioned, is intentional on the customer's part to carry some safety or buffer stock in anticipation of limited supply next year. The naturally economy-sensitive consumer electronics makers are likely to continue remaining conservative in their procurement patterns and carry lean levels of inventory until they see actual positive signals towards demand recovering but that means we could even expect some upside when the end markets do actually improve because these consumer electronics makers first could convert or transition towards more active procurement.

[Foreign Language] The next question will be presented by Nicolas Gaudois from UBS.

Just following up a little bit on the latter question. You're guiding for Q2 bits to be up sequentially double-digit for both DRAM and NAND. Would you characterize this as being driven by end demand actually starting to recover in some segments and/or customers actually starting to restock? And if so, where? And second question, post-Samsung announcement on production cuts earlier this month, are you revising further the degree to which you are lowering production on your side? Are you making incremental wafer start adjustments? If so, could you give us more color on the magnitude duration of its cuts and for which products within the DRAM and NAND.

[Foreign Language]

[Interpreted] I will take your first question on the bit shipment growth outlook for Q2. In the first quarter, as you know, both DRAM and NAND bit shipments declined substantially due to weak seasonality as well as weakness across end markets while our customers continue to digest inventory. We believe that the upside we are seeing for Q2 is mostly driven by inventory buildup demand from the customers ahead of the second half because we're currently seeing significantly broad based our customer inventory levels close to peak. But the situation, of course, depends on the actual application. We're currently seeing real demand. We're receiving real demand from our customers for the high-density DDR5 offering as well as high-density mobile and graphics products. But as you would know, for computing application phase DDR4, this product is very highly stopped at the moment. So we're seeing real demand trending very weak in Q2 for DDR4. To add the bit growth decline in Q1 was a result of ongoing inventory correction, and we believe the magnitude was larger than what would have been driven by just demand. So we believe that the growth in Q2 would also actually represent a normalization of this trend compared to the prior quarter.

Now on your question about our production adjustment in line with the demand environment. The memory industry currently faces a supply-demand mismatch as well as inventory levels that are unprecedented in scale and their magnitude in recent history. Due to these grave market circumstances, we are seeing DRAM a longtime profit generator also turn losses in Q1 across the industry, across DRAM players and the losses widening in the case of NAND Flash business. And given the demand landscape, we believe that it would not be likely that -- any more likely that we can expect a dramatic increase in ASPs in Q2 either. In light of these circumstances, SK hynix is taking a flexible and responsive approach to our production operations and are adjusting our wafer starts, mainly for the products that are currently heavily stocked. As the effect of production cuts of the industry began to materialize in Q2 and on top of that, as the more recent production adjustment aligned to the demand landscape come into play, we expect the supply-demand situation, along with the market trends to improve beginning Q3. The company plans to maintain its conservative production plan until we see a more balanced supply/demand situation as well as more adequate inventory levels.

[Foreign Language] The next question will be presented by Giuni Lee from Goldman Sachs.

[Interpreted] I have two questions, one on AI servers; second on cash flow. First, I want to know the company's outlook on the AI server-related memory market, how are you preparing product portfolio wise and what are your broader plans? Secondly, I understand that you successfully financed yourself in Q1 by issuing [ 1 ] and dollar-denominated bonds, so I want to know more about the exact cash flow that you reported in Q1.

[Interpreted] I'll take your question on the memory outlook in relation to AI servers. While the AI server -- the broader AI market outlook, could range very widely depending on the exact workload that the AI server or system targets hynix, SK hynix is currently focusing on the best-case scenario for the AI market growth. And on top of that, we believe that the market is very receptive to such positive upward change or growth because the value experience that you've gained from AI is already very prevalent, so we believe that growth will be stronger incrementally going forward. To give you the numbers, we believe that adding maximum server shipments and related memory shipment growth could amount to nearly 40% over the next 5 years, whereas in revenue terms for DRAM and NAND combined, that could represent -- AI could represent a 30% growth opportunity for the next 5 years. For the immediate term, we are seeing a positive impact to our business from the DDR5 128 gigabyte or higher density modules as well as HBM offerings as was mentioned in the prepared remarks. Revenue-wise, we expect, again, that DDR5-based high-capacity modules would grow in revenue more than sixfold over year '22 and HBM would see a more than 50% growth year-over-year. And we understand that most of these orders are actual orders that have been received. And furthermore, we expect even stronger growth next year for both the DDR5 and HBM offerings. So for our product competitive edge, we currently have in our line up a 128-gigabyte server DIMM that is built on DDR5. But later on, we are planning higher module capacities are going up to 256 gigabytes for the immediate term based on 32 gigabit mono die density DDR5. And then for the HBM offerings, we are later planning an 8-gigabit per second rated HBM3E to follow up with our current premium offering. Now on first quarter cash flow. So on cash flow, our cash flow from operating activities was a negative KRW 2 trillion in the first quarter as you have heard in our earnings report. In the first quarter, EBITDA declined significantly, as you may have recognized from the earnings report due to a significant decline in our top line as well as an increase in our operating loss. On top of that, we saw working capital increase as inventory grew, and we also had to pay out profit sharing bonuses to our employees based on last year's performance. And then, we also had to reflect the more than KRW 3 trillion in PP&E acquisition costs, including payables carried over for equipment that were installed last year. In the following quarters, we expect the cash outflow to be limited as we are moving in line with our announced investment cuts. And further on, as revenue grows and -- excuse me, inventory is worked down, we expect there to be more cash inflows. So net for net, we would say that after Q2, we would see a more stable management of our overall expenses as well as cash flow.

[Foreign Language] The next question will be presented by J.J. Park from JPMorgan.

[Interpreted] My first question is following your competitors' production cut announcement, I want to ask the company if you see this as reflecting an actual change in customer sentiment because, for example, in the spot market, you are seeing spot prices plateauing or leveling out after significant declines to date following the announcement. And then secondly, this question is about your current year interest expense as well as your portion of debt or amount of debt set to mature in year '23 as well as 24.

[Interpreted] I'll take your first question. You've asked whether this production cut announcement induced spot market movement and other phenomena are due to a change in consumer sentiment. But whether that is the case, whether we are seeing actual customers sentiment change, we cannot say for sure that is the cause, but we do have some explanation to offer. So we can cite the actual questions we're getting from customers. One, some customers are asking if they should now begin preparing for demand returning in the second half by building up their inventory, i.e., resuming procurement starting Q2. And then secondly, they're asking whether the spot price has bottomed and whether that will begin translating into a contract price upside and stabilization. Specifically, our customers are most interested and keen to know whether they will get stable supply of particular products such as the high-performance and high-density LPDDR5 for mobile as well as server DDR5 and our graphics products. So related discussion have actually begun around this rising interest. Then about your question on our interest expense, expectation for this year following our new borrowings as well as the expectation for debt maturing. You would know well about the financial market conditions that continue to persist. Likewise, we also expect high interest rates across the world to persist in the year '23. Following the new round of debt financing, we expect interest expense for year '23 to double roughly over year '22 to KRW 1 trillion. Over the next 3 years, we expect an annual average KRW 4 trillion to KRW 5 trillion in debt to reach maturity. As mentioned in the remarks, our plan is to maintain current investment and cost efficiencies in the following upturn to long term, gradually work down the size of our borrowings.

[Foreign Language] The next question will be presented by Sung Kyu Kim from Daiwa Securities Korea.

[Interpreted] First, following the U.S. government's announcement of the guardrail provision to the CHIPS and Science Act. Now it is the case that in China, you cannot add or expand new capacity for your advanced processes but it seems that you can continue on with process migrations. So what specifically are the company's China fab plans in light of these developments? And secondly, about your inventory write-down charges or inventory valuation loss. What exact number did you see in Q1? And what's your outlook for Q2?

[Interpreted] Thank you. First, on the process migrations in China and our plans. Currently, we are considering various factors, including the long-term geopolitical risk market demand and fab operation efficiencies in determining our direction for China fabs going forward, and we are looking at various different scenarios, considering these. However, nothing is set in stone yet or has significantly changed in our as-is operation plans for our China facilities. Our current focus, of course, is on maintaining the stability of our China fab operations, and we are also positive towards our likelihood of receiving an extension for the U.S. government's waiver on equipment export controls into China. Then on the inventory valuation loss. In the first quarter, the company recognized close to KRW 1 trillion in inventory valuation loss as inventory grew while ASP declined quarter-on-quarter. However, if there are any inventory write-down charges that do occur in the second quarter, we believe they would be significantly smaller, if any, because we expect inventory to peak in Q2 as well as for the ASP declines to moderate significantly. In the second half, we expect inventories to gradually decline and ASPs to stabilize, in which case we could even expect a reversal in the losses that we saw prior as we sell undervalued inventory, and this would be contributional or contributing to our P&L recovery.

[Foreign Language] The next question will be presented by Hyunwoo Doh from NH Securities.

[Interpreted] My first question is on your DDR5. Your DDR5 seems to be more competitive in terms of product quality compared to your peers. How is this so from the company's viewpoint? And also, what's the current profit margin difference of DDR5 over DDR4? Secondly, I have also heard about the company's leadership very much in HBM. So could you provide more color on your business outlook for the HBM segment? And also likewise in DDR5, how are you more competitive than your peers in HBM?

[Interpreted] Thank you. First, on DDR5's competitive edge. Okay. Back in 2018, we were the world's first to develop, successfully develop a 16-gigabit DDR5. And after that followed up with the 24-gigabit DDR5 that sampled with customers in 2021. Then just last year, we delivered samples of a 6,400 megabit per second rated DDR5 modules again as the world's first. So I would say that we have leadership in the development and the delivery to market of the DDR5 segment. That allowed us to gain priority in securing the necessary module components and also to ensure product quality of the components that we sourced that were necessary for these DDR5 modules. On top of that, we were able to work very closely in advance with our partners and customers in verifying the reliability of these modules along the development process from very early on. So we believe this drove our competitive advantage indeed in DDR5. Then on product supply side, DDR5 crossover compared to DDR4 is expected for the server segment in Q2 '24 whereas we expect the timing of bid crossover for DDR5 to the Q1 '24 for the PC segment. But internally, we are minding and working towards an even earlier time line to transition to DDR5 in our DRAM lineup within the second half. So we believe that we are quite prepared and working with great focus to pursue our DDR5 leadership further. Then on the margin difference of DDR5 over DDR4. Well, it would be difficult to compare apples-to-apples because the cost structure is very different between DDR5 and D4 because DDR5 has many BOM cost, bill of material cost adders, minding that though, we are operating our market price to minimize that gap that prevents DDR5 purchasing due to that cost difference. Then on the HBM business outlook and our competitive edge. I could summarize our HBM advantage in 4 different points. One would be our stellar performance of HBM products driven by robust designs. Secondly, industry-leading quality Thirdly, our proactive investment into HBM over the years that ensured fast time-to-market and sufficient capacity. And then lastly, our strong collaboration with SoC or system on chip, chipset makers across the world as well as cloud service providers who are our future customers in the HBM field. More specifically on the performance side, we currently market an HBM3 that is rated at 5.6 gigabits per second in data rate, but we plan to sample an 8 gigabit per second, HBM3E, in the second half of this year, targeting volume production ramp in the first half 2024. On the quality side, the most crucial issue would be to prevent excessive heat that can be generated in the process of stacking these individual HBM layers and connecting them using through silicon via interconnects. So from HBM2E onwards, we have used very unique and proprietary technology in order to ensure the industry's best heat dissipation of quality for our HBM products. On the capacity side, we are preparing HBM production capacity that is enough to support more than 50% annual growth going forward, and we are translating this into actual business opportunity with our customers. Finally, while the main HBM customers used to be chipset makers, that provide various system on chips, including AI or accelerator cards, we expect cloud service providers to also become a major customer segment going into the future, so we are engaging in actual collaboration with the CSPs already and believe this is a great opportunity to expand in the HBM segment going forward.

[Foreign Language] The last question will be presented by Sunwoo Kim from Meritz.

[Interpreted] I have two questions. First, on your NAND Flash business. We are now seeing your NAND profit margins having declined significantly to a very worrying levels. While in the past year strategy and the industry strategy was to scale up simply, now we believe it is the time to focus on profitability improvement. What kind of strategic plan what the company have for your NAND Flash profit margin? And then in Q4 in relation to this, I believe there was an additional cost coming from the Dalian fab, amounting to KRW 200 billion that caused a revision to your reported performance. I wonder if this was somehow reflected in your operating loss or was this also a nonrecurring cost. Secondly, as you mentioned, HBM, I do agree that you are doing very great preparations across performance, quality, time-to-market and capacity. But I have a suggestion and question because HBM also has very high BOM cost, and if it were to follow the business model of commodity memory, then it will be very dependent on the cycles. So I wonder if the company has any plans to turn this into a made-to-order model perhaps in order to secure more visibility in terms of price and quantity.

[Interpreted] Thank you. I'll take your question on NAND. Our NAND business for many years before remained at roughly a 10% share of our business. So a relatively small share. And we also saw a significant technology and product portfolio gap with competitors during those years. To make up for this gap and rise in our NAND business, we took 3 measures. First, to scale up the business, and now we have reached economy of scale. Secondly, we have reached a higher mix of production -- higher mix of bits for our 176-layer stack for the NAND Flash, which is higher than the competition, and this allowed us to catch up in terms of the technology road map and also the cost competitiveness. Thirdly, we pursued an M&A with another NAND Flash player in the market in anticipation of potential restructuring of the industry over the long term. As a result of such efforts, we were able to raise our combined NAND Flash market share to the 20% level. And also product portfolio-wise, we diversified from a mobile focus to one that includes a broader range of SSDs. That said, year '22 being the first year since the Solidigm acquisition, there were various standup costs as well as charges from acquisition accounting, including the purchase price allocation that added to the nonrecurring costs that impacted performance. While we have secured a product level advantage by securing and achieving a higher mix of 176-layer bits, we expect our business -- NAND business to face a challenging outlook for the immediate future given the continued market weakness as well as continued cost from the M&A/PMI procedure. We are currently laser-focused on reducing redundant and inefficient cost, cost-driving matters, such as by implementing company-wide CapEx and OpEx controls and also continuing to integrate our 2 companies' individual functions and overall streamlining our cost structure. So when business returns, we will be completely ready to convert to a mix of premium high value-added products such as high spec and advanced enterprise SSDs so as to ensure a quick rebound. Then about your second question. I believe that the costs, the added cost coming from the Dalian fab was a result of the difference in the timing of the performance report and then the timing when the costs were closed and reported. Due to the time lag of the company's report and also the separate and consolidated financial reporting, the Dalian fab and related costs also have not been reflected into our full performance just yet due to the time lag in reporting. And then on your third question. First, I would like to note my gratitude for your interest and notice of our technology advantage in HBM3 and DDR5. For the HBM market, which is benefiting from the proliferation of AI services like ChatGPT, we have very -- a handful of customers that can actually build the AI compute-oriented high-end graphic systems and even fewer suppliers that can provide them with the latest and fastest HBM3. Given this limited customer and supplier base, we could consider using this as an opportunity to transition to a pricing and business model that can deliver greater value through customer growth and more stable and supply visibility, excuse me. If we are able to implement such business model, we expect there to be benefits, including a smoothing out of the profitability ups and downs as well as greater visibility into our business growth.

[Foreign Language] [Portions of this transcript that are marked [Interpreted] were spoken by an interpreter present on the live call.]